EMISSIONS CONTROL OPTIONS
FOR THE SYNTHETIC ORGANIC CHEMICALS
MANUFACTURING INDUSTRY
Fugitive Emissions Report
February 1979
-------
DRAFT
EMISSIONS CONTROL OPTIONS FOR THE
SYNTHETIC ORGANIC CHEMICALS MANUFACTURING INDUSTRY
EPA Contract No. 68-02-2577
Fugitive Emissions Report
D. G. Erikson
V. Kalcevic
Prepared for
Emission Standards and Engineering Division
Office of Air Quality Planning and Standards
ENVIRONMENTAL PROTECTION AGENCY
Research Triangle Park, North Carolina
February 1979
-------
iii
CONTENTS
Page
I. PREFATORY AND INTRODUCTORY MATERIAL 1-1
II. CHARACTERIZATION AND DESCRIPTION II-l
A. Introduction II-l
B. Sources II-2
C. Methodology II-9
D. References 11-15
III. CONTROL TECHNOLOGY III-l
A. Control Devices III-l
B. Leak Detection Methods III-5
C. Maintenance III-1Q
D. References 111-12
IV. ANALYSIS IV-1
A. Introduction IV-1
B. Control Costs IV-3
C. Control Impact IV-16
D. References IV-18
V. ASSESSMENT V-l
A. Summary V-l
B. Supplemental Information V-l
C. References V-3
-------
V
APPENDICES
Page
APPENDIX A. TOTAL NUMBER OF SOCHI PRODUCT SITE LOCATIONS A-l
APPENDIX B. MONITORING/MAINTENANCE MANPOWER REQUIREMENTS FOR FOUR B-l
MONITORING PROGRAMS
APPENDIX C. NUMBER OF PUMPS AND VALVES VERSUS PLANT CAPACITY AND C-l
TYPICAL MODEL PLANT
APPENDIX D. COST ESTIMATE DETAILS D-l
APPENDIX E. LIST OF EPA INFORMATION SOURCES E-l
-------
vii
TABLES
Number Page
II-l SOCMI Valve Characterization 11-10
II-2 SOCMI Pump Seal Characterization 11-10
II-3 Number of Equipment Components by Chemical(s) 11-12
II-4 Current Total SOCMI Equipment Component Estimate 11-14
IV-1 Technical Parameters for Three Model Plants IV-2
IV-2 Cost Factors Used in Computing Annual Costs for Equipment IV-4
Modifications
IV-3 Cost Estimates for Installation of Control Devices in a Small IV-5
Model Plant
IV-4 Cost Estimates for Installation of Control Devices in a Medium IV-6
Model Plant
IV-5 Cost Estimates for Installation of Control Devices in a Large IV-7
Model Plant
IV-6 Factors Used in Computing Annual Costs for Monitoring Programs IV-10
IV-7 Control Cost Estimates of Monitoring and Maintenance Program A IV-11
for Three Model SOCMI Plants
IV-8 Control Cost Estimates of Monitoring and Maintenance Program B IV-12
for Three Model Plants
IV-9 Control Cost Estimates of Monitoring and Maintenance Program C IV-13
for Three Model Plants
IV-10 Control Cost Estimates of Monitoring and Maintenance Program D IV-14
for Three Model Plants
A-l Total Number of SOCMI Product Site Locations A-3
B-l Basis for Determining Monitoring/Maintenance Manpower B-4
Requirements
B-2 Annual Monitoring Manpower Requirements for Program A for B-5
Three Model Plants
B-3 Annual Monitoring Manpower Requirements for Program B for B-6
Three Model Plants
B-4 Monitoring Manpower Requirements for One Unit Area Survey B-8
B-5 Annual Monitoring Manpower Requirements for Program C for B-9
Three Model Plants
B-6 Annual Monitoring Manpower Requirements for Program D for B-10
Three Model Plants
B-7 Annual Maintenance Manpower Requirements for Three Model Plants B-12
C-l Model Plants Characterized to Date C-3
-------
ix
FIGURES
Number
Page
II-l
Diagram of a Simple Packed Seal
II-3
II-2
Diagram of a Basic Single Mechanical Seal
II-3
II-3
Diagram of a Double Mechanical Seal
II-3
II-4
Diagram of a Gate Valve
II-5
II-5
Diagram of a Spring-Loaded Relief Valve
II-6
II-6
Liquid-Film Compressor Shaft Seal
II-8
III-l
Diagram of a Rupture Disk Installation Upstream of a Relief Valve
III-3
III-2
Diagram of Simplified Closed-Vent System with Dual Flares
III-4
III-3
Diagram of Two Closed-Loop Sampling Systems
III-6
C-l
Plot of Total Number of Pumps per Plant vs Plant Capacity
C-4
C-l
Plot of Total Number of Valves per Plant vs Plant Capacity
C-5
-------
xi
ABBREVIATIONS AND CONVERSION FACTORS
EPA policy is to express all measurements used in agency documents in metric
units. Listed below are the International System of Units (SI) abbreviations
and conversion factors for this report.
To Convert From
To
Multiply
Pascal (Pa)
Atmosphere (760 mm Hg)
9.870 X 10~6
Joule (J)
British thermal unit (Btu)
9.480 X 10~4
Degree Celsius (°C)
Degree Fahrenheit (°F)
(°C X 9/5) +
Meter (m)
Feet (ft)
3.28
Cubic meter (m3)
Cubic feet (ft3)
3.531 X 101
Cubic meter (in3)
Barrel (oil) (bbl)
6.290
Cubic meter (m3)
Gallon (U.S. liquid) (gal)
2.643 X 102
Cubic meter/second
Gallon (U.S. liquid/min)
1.585 X 104
(m3/s)
(gpm)
Watt (W)
Horsepower (electric) (hp)
1.340 X 10~3
Meter (m)
Inch (in.)
3.937 X 101
Pascal (Pa)
Pound-force/inch2 (psi)
1.450 X ID'4
Kilogram (kg)
Pound-mass (lb)
2.205
Joule (J)
Watt-hour (wh)
2.778 X 10~4
Standard Conditions
68°F = 2Q°C
1 atmosphere = 101,325 Pascals
PREFIXES
Multiplication
Prefix Symbol Factor Example
T tera 1012 1 Tg = 1 X 1012 grams
G giga 109 1 Gg = 1 X 109 grams
M mega 10s 1 Mg = 1 X 10s grams
k kilo 103 1 km = 1 X 103 meters
m milli 10~3 1 mV = 1 X 10~3 volt
)j micro 10 6 1 pg = 1 K 10 6 gram
-------
(This section to be
i-l
I. PREFATORY AND
supplied by EPA.)
INTRODUCTORY MATERIAL
-------
II-l
II. CHARACTERIZATION AND DESCRIPTION
A. INTRODUCTION
The Synthetic Organic Chemicals Manufacturing Industry (SOCMI) is a segment of
the domestic chemical industry that produces 350 to 400 basic and intermediate
organic chemicals used to produce other intermediates and finished chemicals.
Organic chemicals not included in the SOCMI are refinery by-products, coal-tar
products and other "naturally" derived organic chemicals or polymers.
For the purpose of this report fugitive emissions are defined as those volatile
organic compound (VOC) emissions that result from leaking plant equipment. The
potential sources of VOC emissions in the SOCMI that are considered in this
report include pump seals, in-line (process and control) valves, relief valves,
open-ended (sample, drain, vent) valves, compressor seals, flanges, and cooling
towers. VOC emissions that result from the transfer, storage, treatment, and
disposal of liquid, solid, and aqueous process wastes are defined as secondary
emissions and will be covered in a separate report.
The data base that was used for this report to characterize the industrial types
of equipment was compiled from 53 plant site visits1 and 2 EPA reports: Equipment
Component Analysis for Identification of Potential Fugitive Emission Sources,2
prepared by Pullman Kellogg, and Data Package for Formaldehyde Plant Fugitive
Emissions Study,3 prepared by Walk, Haydel and Associates, Inc.
The data supplied from the site visits and the two EPA reports included a quanti-
fication of the number of equipment components associated with a specific chemi-
cal or co-products manufactured from a specific process. The various equipment
components, all of which are considered to be potential sources of fugitive
emissions, were grouped into five major equipment categories: pumps, valves,
compressor seals, flanges, and cooling towers. Pumps were further subdivided
as to the type of seal used, with four subcategories considered: those with
single mechanical seals, double mechanical seals, packed seals, or no seals
(sealless). Valves were subcategorized by type, with three groupings considered:
saftey/relief, open-ended, and in-line. These three valve groupings were further
subdivided by the kind of service: vapor (gas) or liquid.
-------
II-2
B. SOURCES
There are many kinds of equipment in the SOCHI that can contribute VOC emissions
from leakages. The types of equipment assessed for fugitive emissions in this
report are as follows.
1. Pumps
Pumps are used extensively in the SOCHI for the movement of organic fluids.
The predominant type used is the centrifugal pump, although other types, such as
the positive-displacement pump, both reciprocating and rotary action, and the
specialized canned and diaphragm pumps, are used for some applications. Except
for such pumps as the canned and diaphragm types, the pumps have a shaft that
requires a seal to isolate the pump interior fluid from the atmosphere. The
possibility of a leak through this seal makes it a potential source of fugitive
VOC emissions.
Three general types of shaft seals are in use, packed, single mechanical, and
double mechanical, which are listed in order of increased effectiveness in mini-
mizing leaks. The sealless pump is, of course, the most effective. Proper instal-
lation and maintenance are required for all seal types if they are to function
properly and retain their ability to seal.
a.. Packed Seal — Figure II-l is a diagram of a simple packed seal. Packed seals
can be used on both reciprocating and rotary action types of pumps. The seal
consists of a. stuffing box in the pump casing filled with specialized packing
material that is compressed with a packing gland to form a seal around the shaft.
To prevent buildup of frictional heat, lubrication is required. K sufficient
amount of either the liquid being pumped or of a liquid that is injected must
be allowed to flow between the packing and the shaft to provide the necessary
lubrication.
b. Single Mechanical Seal — Figure II-2 is a diagram of a basic single mechanical
seal. Mechanical seals can be used on rotary-type pumps only. The rotating-
seal-ring face and the stationary-element face are lapped to a very high degree
of flatness to maintain contact throughout their entire mutual surface area.
As with packing, the faces must be lubricated; however, because of its construc-
tion, much less lubrication is needed. There are many variations to the basic
-------
II-3
FUUIO
EJ>JO
WWP
STUFVlMft
bOK
£n
^PACXIKJi
GUAJJO
S^EAiJfXCE/^ _ J
Ixixixjxixixi^iz:
i>
PosaibLC.
LtAK.
AHEX
- PAOCJ VJGi
Fig. II-l. Diagram of a Simple Packed Seal
PUMP
4TUFFIU&
BOX.
FCUIO
ElUO
UUO
R.IM&
VJSER.T PACKJlJtm
aTATIOUARY
E.LE.MEJJT
POSSI6L.E
|_C-A>K.
AREA
Fig. II-2. Diagram of a Basic Single Mechanical Seal
Fig. II-3. Diagram of a Double Mechanical' Seal
-------
II-4
design but all have the lapped seal face between a stationary element and a
rotating seal ring.
Double Mechanical Seal — Figure II-3 is a diagram of a double mechanical seal
in which the two seals are installed in opposite configuration. A liquid, such
as water or seal oil, is circulated through the stuffing box space between the
two seals. This arrangement is an improvement over a single seal since the
sealing fluid surrounds the double seal and provides lubrication to both sets
of seal faces. It also allows a low differential pressure to be set across the
seal face.
Another double mechanical seal arrangement is one in which the ttfo seals are
installed in the same, or tandem, configuration. The inner seal functions identic-
ally to a conventional single inside seal. The stuffing box space between the
two seals is flooded with a liquid from a closed reservoir. If the inner seal
fails, it will be sensed by a pressure rise at the reservoir,- also, the outer
seal will take over as a backup seal.
Valves
One of the most common pieces of equipment in organic chemical plants is the
valve. The types of valves commonly used are control, globe, gate, plug, ball,
relief, and check valves. All. except the relief valve and check valve are acti-
vated by a valve stem, whose motion may be rotational or linear or both, depending
on the specific design. This stem requires a seal to isolate the valve interior
fluid from the atmosphere. The possibility of a leak through this seal makes
it a potential source of fugitive VOC emissions.
The most common type of valve stem seal in use is the packed seal. It consists
of a stuffing box in the valve housing filled with specialized packing material
that is compressed with a packing gland to form a seal around the stem.
Figure II-44 is a diagram of one type of valve with a valve stem.
Pressure-relieving devices are required by engineering codes for applications
where the pressure on a vessel or a system may exceed the maximum allowed.
Spring-loaded safety/relief valves are typically used for this service in the
SOCHI. Figure II-5 is a diagram of a relief valve. The seal is a disk on a
-------
II-5
Fig. II-4. Diagram of a Gate Valve (from Ref. 4)
-------
n-6
PROCESS SI OS.
Fig. II-5. Diagram of a Spring-Loaded Relief Valve
-------
II—7
seat held in place during normal system operation by a spring. The possibility
of a leak through this seal makes it a potential source of fugitive VOC emissions.
There are two potential causes for leaks: "simmering,11 a condition due to the
system pressure being close to the valve set pressure, and improper reseating
following a relieving operation.
Check valves are used to prevent fluid backflow in a system. They do not need
a seal and therefore are not a source of leakage.
Some valves are installed in a system so that they function with the downstream
side open to the atmosphere. Examples are sample valves, drain valves, and
vent valves. The possibility of a leak through the seat of these valves makes
them a potential source of fugitive VOC emissions.
3. Flanges
Flanges are bolted gasket-sealed junctions used wherever pipe or equipment com-
ponents such as vessels, pumps, valves, and heat exchangers may require isolation
or removal. The possibility of a leak through the gasket seal makes them a
potential source of fugitive VOC emissions.
Two primary causes of leakage is seal deformation due to thermal stress on the
adjoining piping or equipment and repeated opening without replacement of the
gasket.
4. Compressors
SOCMI compressors, like pumps, can be both centrifugal and positive displacement
types. Compressors have a shaft that requires a seal to isolate the compressor
interior gas from the atmosphere. The possibility of a leak through this seal
makes it a potential source of fugitive VOC emissions. In addition to having
seal types like those for pumps, centrifugal compressors can be equipped with a
liquid-film seal as shown in Fig. II-6.5 The seal is a film of oil that flows
between the rotating shaft and the stationary gland. The oil that leaves the
compressor from the system side is under the system internal gas pressure and
is contaminated with the gas. When this contaminated oil is returned to the
oil reservoir, process gas can be released and emitted to the atmosphere.
-------
II-8
IUME.R
&U5H1M&
/-OIL. INi Pf?.OM RESERVOIR.
,OUTtt
&USH1WGJ
JklTERSJAU
GAS PRS.SSURS.
COW TAM1 MATEJD
OIL. OUT
TO R.e.ScX.VOlR
ATMOSPHERE
OIL.,OUT
Fig. II-6. Liquid-Film Compressor Shaft Seal (from Ref. 5)
-------
II-9
5. Cooling Towers
Cooling towers cool the recirculating water that is used to remove heat from
such process equipment as reactors, condensers, and heat exchangers. If a leak
in the process equipment occurs and if the equipment is operating at a pressure
higher than that of the recirculating water, process material can get into the
water stream. This material can be released to the atmosphere at the cooling
tower, making it a potential source of fugitive VOC emissions. A potential
source of fugitive emissions can also occur if VOC-contaminated process water
is used as the cooling water source.
6. Other Sources
It is recognized that other kinds of equipment in the SOCMI can Contribute VOC
emissions from leakages that are not covered in this report. These include
threaded connections and agitator seals. The use of threaded connections in
VOC service is believed to be small, accounting for an estimated 2 to 3% of all
piping-related connections.2 Agitator seals could not be assessed for their
fugitive emission potential because of a lack of sufficient data base information.
C. METHODOLOGY
1. Data Base
. The data base was compiled in two steps. In the first step all the equipment
component data supplied by the reference sources for the valve and pump subcate-
gories were averaged to yield a typical characterization for the industry.
This approach was considered to be reasonable since the data represented both
large- and small-capacity processes using both continuous and batch production
methods. Results of the initial characterization of valves and pumps from the
data base are given in Tables II-l and II-2. This information is considered to
be representative of the entire industry.
From Table II-l safety and relief valves constitute 3.4% of the total valve
population. An estimated 83% of the total safety relief valves are in vapor
service and 17% are in liquid service. Open-ended valves, which include sample,
vent, and drain valves, represent 27.6% of the total valve count and find primary
use in liquid service (91%), with a small application (9%) in vapor service.
In-line valves, including process and control valves, make up the bulk of the
-------
11-10
a
Table II-l. SOCMI Valve Characterization
Type of
Valve
Percent of
Total Valves
Percent in
Licfuid Service
Percent
Vapor (Gas)
in
Service
Safety/relief
3.4
17.0
83.0
Open-ended (sample
vent, drain)
27.6
91.0
9.0
In-line (process,
control)b
69.0
65.0
35.0
All valves
100.0
71.0
29.0
includes only valves in VOC service.
^Check valves are excluded.
Table II-2. SOCMI Pimp Seal Characterization*
Punro Seals
Percent in Use
Mechanical
Single
71.9
Double
16.7
Packed
10.6
None (sealless)
0.8
Total
100.0
*Includes only pump seals in VOC service.
-------
11-11
total valve population (69%) in the industry, with 65% of the total in-line
valves in liquid service and 35% in vapor service. In general, valves in the
industry are used primarily in liquid service (71%), the remaining 29% being
used in vapor service.
As shown in Table II-2, the most common type of pump seal in use in the industry
is the mechanical seal. An estimated 88.6% of all pump seals used to handle
VOC are mechanical; 71.9% of them are single mechanical seals and 16.7% are
double mechanical seals. Pumps with packed seals represent 10.6% of the total
pumps. Less than 1.0% of the total industry pumps are sealless.
A basic assumption used in estimating the total SOCMI equipment iomponents is
that the total number of components for each chemical is dictated more by the
number of plants manufacturing the product than by the capacity of the produc-
tion facility. A plant that is twice as large in capacity as another, producing
the same chemical, does not necessarily have twice as many pieces of equipment.
In most instances the equipment is simply larger in size.
In the second step the data obtained from the site visits and the two EPA reports
enabled a quantification to be made of the total number of valves, pumps, and
compressors for each of 35 different chemicals. This information was then scaled
up to estimate the total quantity of valves, pumps, and compressors for all the
domestic production sites for each of the 35 chemicals. Table II-3 lists the
35 chemicals in alphabetical order that are covered by the data base, the number
of sites at which each is produced, and the estimated total number of pumps,
valves, and compressors that are in service at all sites for each chemical.
The totals for valves, pumps, and compressors for the 35 chemicals were summed,
along with the number of sites. The 35 chemicals in the data base constitute
415 product sites in the SOCMI, which contribute an estimated total 14,391 pumps,
358,034 valves, and 590 compressors. These total equipment components were
\
then divided by the total number of sites (415) to yield an average number of
pumps., valves, and compressors per site, i.e., 35 pumps per site, 865 valves
per site, and 1.4 compressors per site.
-------
11-12
Table II-3.
Number of Equipment Components
in Chemical Manufacturing
Used
Number of
Plant Sites
Total Number of Equioment Components'
Chemical(s)
P vinos
Valves
Coooressors
Acataldehyde
4
67
7,098
0
Acetic acid
9
154
6,199
0
Acetic anhydride
6
152
4,000
0
Acetone cyanohydrin
3
48
1,263
0
Acrolain/glycerin
4
636
3,000
0
Acrylic acid
3
132
3,156
0
Acrylic acid eaters
3
369
7,796
0
Acrylonitrile
6
195
7,101
6
Adipic acid
S
180
4,734
0
Aniline/nitrobenzene
7
486
6,629
*0
Benzene
14
62
S.779
31
Butadiene
20
907
33,941
180
Caprolactan
3
674
17.713
0
Chlorobenzenes
e
311
5,657
2
Chloroaethanes
17
S10
17,000
34
Cunene
14
608
12,765
0
Cyelohexane
11
127
3.24S
6
Cyclohexanone/cyclohexanol
8
328
3,626
0
Dimethyl terephthalate
6
392
15,428
3
Ethyl acetaee
3
120
2,140
0
Ethyibenzene/styrene
19
315
3,087
10
Ethylene/propylane-
37
2.095
52,690
185
Ethylene diehloride
17
714
13.778
0
Ethylene oxide
16
144
10,635
29
Formaldehyde
53
S34
11,953
35
Glycol ethers
9
135
2.574
0
Linear allcyl benzene
4
266
10,042
4
Maleic anhydride
9-
138
3,724
10
Methanol
12
360
6,616
12
Methyl rnethacrylate
S
123
3.236
0
Phenol/acetone/methyl 5Cyrano
11
1,376
11,207 .
0
Terephthalic acid
3
121
2,OSS
20
Urea
40
100
2.68S
0
Vinyl acetate
7
312
12.732
18
Vinyl chloride
14
700
22,750
0
Data base total
415
14,391
358.034
590
•In VOC service.
-------
11-13
Flange estimates were made by developing a flange-to-valve ratio from the Pullmann
Kellogg2 and Walk, Haydel reports3 and data obtained from visits to selected
sites. A flange-to-valve ratio of 1.6:1 was considered to be representative of
the data base.
2. Industry Scaleup
From the EPA Report Organic Chemical Producers Data Base Program Volume II,6
prepared by Radian Corp., 1334 unigue product sites were identified with the
production of the 365 organic chemicals by the SOCMI.
The total industry estimate of equipment components was obtained by multiplying
the total number of SOCMI sites (1334) by the average number of Components per
site obtained from the data base. This extrapolation is considered to be valid
because the data base, consisting of 62 plant sites, represents approximately
5% of the total SOCMI plant sites and incorporates both large and small capacities
and both batch and continuous production methods.
The number of process cooling towers for the industry was determined, from the
data base, to average one per production site. An average flow rate 5000 gpm
for the tower was estimated as being representative of the data base and the
industry.
The current, total, SOCMI .equipment component estimate resulting from the scaleup
is shown in Table II-4.
-------
11-14
Table II-4. Currant SOCMI Equipment Component Estimate3
Equipment Individual Total
Component Quantity Quantity
Pumps
With mechanical seals
Single 33,588
Double 7,777
Total 41,365
With packed seals 4,950
Sealless 375
Valves
I.
In-line (process)
Vapor service 278,690
Liquid service 517,570
Total 796,260
Safety/relief 39,240
Open-ended (sample) 318,500
Components
in VOC service .
46,690
1,154,000
Flanges 1,846,200
Compressors . 1,870
Cooling towers 1,334
is
Check valves are not included.
-------
11-15
D. REFERENCES*
1. Site visits conducted and used to quantify various equipment components for the
Fugitive Emissions report; see Appendix E.
2. Pullmann Kellogg, Houston, TX, Equipment Component Analysis for Identification
of Potential Fugitive Emission Sources (data on file at EPA, ESED, Research
Triangle Park, NC) (June 1978).
3. Walk, Haydel and Associates, Inc., New Orleans, LA, Data Package for Formaldehyde
Plant Fugitive Emissions Study (data on file at EPA, ESED, Research Triangle
Park, NC) (June 27, 1978).
4. Emissions from Leaking Valves, Flanges, Pump and Compressor Seals, and Other
Equipment at Oil Refineries, Report No. LE-78-001, State of California Air
Resources Board (April 1978).
5. R. F. Boland et al., Monsanto Research Corp., Screening Study for Miscellaneous
Sources of Hydrocarbon Emissions in Petroleum Refineries, EPA Report No. 450/3-76-041
(December 1976).
6. Radian Corp., Austin, TX, Organic Chemical Producers' Data Base Program. Volume II,
Final Report (data on file at EPA, ESED, Research Triangle Park, NC) (Aug. 5,
1976).
*When a reference number is used at the end of a paragraph or on a heading, it
usually refers to the entire paragraph or material under the heading. When,
however, an additional reference is required for only a certain portion of the
paragraph or captioned material, the earlier reference number may not apply to
that particular portion.
-------
III-l
III. CONTROL TECHNOLOGY
This section describes the existing technology for the control of fugitive emis-
sions in the synthetic organic chemicals manufacturing industry (SOCMI). The
control of equipment leaks (fugitive emissions) can be accomplished not only by
the installation of control devices but also by the use of monitoring (leak
detection) and maintenance procedures.
A. CONTROL DEVICES
In some cases the equipment can be modified to limit VOC emissions from SOCMI
plants. The modifications included in this report that are considered to be
control devices consist of double mechanical seals on pumps, rupture disks and
closed-vent system for safety/relief valves, blinds and closed-loop sampling
for open-ended valves, and closed-vent system for compressor liquid-film seals.
1. Double Mechanical Seals
Double mechanical seals are superior to single mechanical seals and pack seals
in preventing leakage from rotary-type pumps and compressors. By design, double
mechanical seals have a chamber between the two seal faces that either is flushed
with a sealing fluid that allows control of the conditions under which the seal
operates or is flooded with a fluid whose pressure can be monitored for seal
failure. Field screening of double mechanical seals has shown that the VOC
leak occurrence is negligible.1
Mechanical seals, single or double, have limits on their applicability. They
can be used only on shafts with a rotary motion. Also, the maximum service
temperature is usually limited to less than 260°C.2 In spite of their limita-
tions it is estimated that about 90% of the SOCMI pumps handling VOC are equipped
with mechanical seals, with 17% being double mechanical seals (Sect. II). There-
fore double mechanical seals could be used in most new-pump applications and
probably could be retrofitted to replace many of the present single mechanical
seals and some of the packed seals. Where the VOC being handled by the pump
has a low vapor pressure the emission potential is reduced, which could limit
the necessity for double mechanical seals in these applications.
-------
III-2
2. Rupture Disks
A rupture disk can be used upstream of a relief valve so that under normal con-
ditions it seals the system tightly but will break when its set pressure is
exceeded, at which time the relief valve will relieve the pressure. Figure III-l
is a diagram of a rupture disk and relief valve installation. The installation
is arranged to prevent disk fragments from lodging in the valve and preventing
it from being reseated should the disk rupture. It is important that no pressure
be allowed to build in the pocket between the disk and the relief valve; otherwise,
the disk will not function properly. A pressure gauge and bleed valve can be
used to prevent pressure buildup. With the use of a pressure gauge it can be
determined whether the disk is properly sealing the system against leaks or whether
it needs to be replaced.
Use of a rupture disk upstream of a relief valve will eliminate leaks due to
the improper seating and to the simmering of the relief valve. Also, the disk
can protect the relief valve against system materials that could be corrosive
and thereby cause improper seating. It is quite common in industry to install
a rupture disk upstream of a relief valve.3
3. Closed-Vent System
A closed-vent system is used to collect and dispose of gaseous emissions in
many industrial operations, particularly in processes where large volumes of
combustible volatile materials are handled. These gases can result from such
things as plant upsets, which require that the material be vented to prevent
dangerous conditions, from plant startups and shutdowns, from disposal of waste
gas streams, and from equipment leaks. Such emissions are typically intermittent,
and their flow rates during major upsets can be large. The usual method of
disposing of them is by flaring. Figure III-2 is a diagram of a dual-flare
system. The smaller flame operates more efficiently with routine smaller exhausts.
The larger flare is normally on standby to handle large emergency exhausts.
By connecting relief-valve discharges and compressor seal oil degassing vents
to a closed-vent system, their emissions can be effectively controlled. The
effectiveness of VOC destruction will depend on the flare design and turn-down
capability. A dual-flare system should be more effective than a single flare
for the relatively low flows from relief valve leaks and compressor seal oil
degassing. A well-designed system can achieve 98 to 99% VOC destruction.4
-------
III-3
BUKJD FL^MGE. REL-tEF VALVE. ATTACHES WE.RE1
Fig. III-l. Diagram of a Rupture Disk Installation Upstream of a Relief Valve
-------
III-4
4 COMPRESSOR. 5£AL OIU
MG V5.UTi
Fig. III-2. Diagram of Simplified Closed Vent System With Dual Flares
-------
III-5
4. Blinds, Caps, and Plugs
Blinds, caps, and plugs are devices for closing off the ends of valves and pipes.
When installed downstream of an open-ended valve, they are effective in preventing
leaks through the seat of the valve from reaching the atmosphere. Open-ended
valves in SOCMI, about 28% of the total valves handling VOC, are used mostly in
intermittent service for sampling, venting, or draining. If a blind, cap, or
plug is used downstream of a valve when it is not in use, VOC emissions can be
reduced. No estimate has been made for their control efficiency. However, the
control efficiency will depend on such factors as frequency of valve use, valve
seat leakage, and material that may be trapped in the pocket between the valve
and blind and lost on removal of the blind. The installation of a blind, cap,
or plug does not prevent the leakage that may occur through the valve stem seal.
5. Closed-Loop Sampling
A frequent operation in many SOCMI plants is to withdraw a sample of material
from the process for analysis. To ensure that the sample is representative,
purging of the sample lines and/or sample container is often required. If this
purging is done to the atmosphere or to open drains and if there are incidental
handling losses, VOC emissions can result. A closed-loop sampling system is
designed so that the purged VOC is returned to the system or sent to a closed
disposal system and so that the handling losses are minimized. Figure III-35
gives two examples of closed-loop sampling systems where the purged VOC is flushed
from a point of higher pressure to one of lower pressure in the system and where
sample-line dead space is minimized. Reduction of emissions from the use of
closed-loop sampling is dependent on many highly variable factors, such as fre-
quency of sampling and amount of purge required; therefore the efficiency of
this control methis is not estimated.
B. LEAK DETECTION METHODS (MONITORING)
In the EPA and industry tests a significant leak is defined as one having a VOC
concentration of over 10,000 ppm at the potential leak source. The emission
rate predicted by linear regression analysis for 10,000 ppm at 0 cm is
1.11 kg/day for pump seals and 0.19 kg/day for valves.6
Leak detection can be accomplished by three types of monitoring: individual-
component survey, unit area survey, and multiple fixed-point monitoring systems.
-------
III-6
PROCESS LIME
=3-
M
a
SAMPLE.
COWTA1NJE.R.
PROCESS LI WE.
t
SAMPLE.
COSJTMUER
Fig. III-3. Diagram of Two Closed-Loop Sampling Systems (from Ref. 5)
-------
III-7
Monitoring methods, their advantages and disadvantages, and their leak detection
effectiveness are discussed.
1. Individual-Component Survey7
Each type of equipment listed in Sect. II can be monitored for leaks by using a
portable VOC detection instrument to sample the ambient air in proximity to the
potential leak point. Both the instrument and the monitoring technique for
each type of equipment are described in Appendix B.
In a complete individual-component survey each leak source is screened by using
a portable VOC detection device to measure the concentration at the surfaces
where leakage could occur. The instrument probe should traverse%the length of
the leak surface to within 1 cm of the surface. Some potential sources, such
as process drains, cooling towers, pressure relief devices, and open-ended valves
or pipes, have an exhaust area open to the atmosphere rather than a seal inter-
face. For these sources the probe should be placed at the centroid of this
open area, as well as around the perimeter of the open area. The major advantage
of the complete individual-component survey is that all significant equipment
leaks are located. By checking each component individually at its surface there
should be no false indications of high concentrations from another leak source.
2. Unit Area Survey7
A unit a'rea survey entails measuring the ambient VOC concentration within a
given distance, for example, 1 m, of all equipment located on ground and other
accessible levels within a processing area. These measurements are performed
with a portable VOC detection instrument utilizing a strip chart recorder. The
instrument operator walks a predetermined path to assure total available coverage
of a unit on both the upwind and downwind sides of the equipment, noting on the
chart record the location in the unit where any elevated VOC concentrations are
detected. If an elevated VOC concentration is recorded, the components in that
area will have to be screened individually to locate the specific leaking equipment.
It is estimated that 50% of all significant leaks in the unit are detected by
the walk-through survey,8 provided that all major leaking equipment has been
repaired, which reduces the VOC background concentration sufficiently to allow
for reliable detection.
-------
III-8
The major advantages of the unit area survey are that significant leaks from
ground and other accessible leak sources can be located quickly and that the
manpower requirements are much lower than those for the individual-component
survey. Some of the shortcomings of this method are that VOC emissions from
adjacent units can cause false leak indications; high or intermittent winds
(local meteorological conditions) can increase dispersion of VOC, causing leaks
to go undetected; and additional effort is necessary to locate the specific
leaking equipment. However, two or more consecutive (back-to-back) surveys in
a unit would minimize these problems. Both the instrument and the monitoring
time requirements are given in Appendix B.
3. Fixed-Point Monitoring Systems7
The basic concept of the fixed-point monitoring system is that sampling-point
devices can be installed at specific sites within the process area to monitor
for leaks automatically. The ambient VOC concentration can be remotely and
centrally indicated to the operator, who can respond appropriately when elevated
levels are recorded. The monitoring sites would not include the entire geographic
area of the facility, but only areas where equipment handling VOC is located.
The performance of individual equipment can also be monitored to detect equip-
ment failures that result in leaks.
The approaches to leak detection with fixed-point monitors differ in the number
and placement, of the sample points and in the manner in which the sample is
taken and- analyzed. One approach is. to establish the sample points near specific
pieces of equipment, such as process pumps, compressors, and cooling towers. A
second approach is to establish the sample points in a grid pattern throughout
the process area. When an elevated concentration is noted, the operator performs
an individual-component survey on equipment in that area to locate the leaking
component. Zn addition to these variations in the location of the sampling
points, different types of systems can be used. For example, the sampling can
be done continuously and the analysis done on-site or the samples can be col-
lected at the site and then analyzed at a central location (an automatic sequential
system).
One feature of the fixed-point monitor approach is that the location of the
monitor and the type of sampling and analysis can be tailored to meet the spe-
cific requirements of individual plant sites and VOC. Fixed-point monitors
-------
Ill—9
have the capability to sample for specific compounds by flame ionization—gas
chromatography (GC/FID) or infrared (IR) analysis. Of the several leak detection
methods, fixed-point monitoring systems have the highest capital cost but the
lowest monitoring manpower requirement. However, this approach may still require
the use of a portable VOC detection device to locate the leak, particularly if
process-area monitoring is used. Leak detection efficiency for fixed-point moni-
toring systems is estimated to be 33%7 for facilities with a small number of recur-
ring leaks, provided that major leaking equipment has been repaired, which reduces
the VOC background concentration sufficiently to allow for reliable detection.
4. Cooling Tower Total Organic Carbon Analyzer
If there is a leak into a cooling water system, detection by a portable or fixed-
point monitor may not be possible at the cooling tower. The air in close proximity
to the tower is normally very turbulent as the result of wind drafts created by
operation of the tower. This can increase the dispersion of VOC and cause the
leak to go undetected. To overcome this problem, a total organic carbon (TOC)
analyzer can be used to monitor for total organic carbon (ppm) in the cooling
tower stream. Total carbon measurements of ^5 ppm indicate a leak.9 Samples
are then taken from various individual items of equipment that make up the cooling
water system until the leak is found. The leak detection efficiency for TOC
monitoring of cooling water with concentrations §5 ppm is expected to be high.
5. Visual Inspection
Visual inspections can be' performed to detect evidence of liquid leakage from
plant equipment. When such evidence is observed, the operator should use a port-
able VOC detection instrument to measure the VOC concentration of the leak. All
liquid leaks will not necessarily result in a reading of greater than 10,000 ppm.6
6. Combined Leak Detection Programs
In actually conducting leak detection programs, combinations of the various leak
detection methods would probably be incorporated. The following four combinations
have been selected for monitoring requirements and cost analysis:
Program A — Individual-component survey with visual inspection as outlined
in Appendix B.
Program B — Fixed-point monitoring with an individual-component survey conducted
initially as outlined in Appendix B.
-------
111-10
Program C — Unit area survey with individual-component survey conducted initially
as outlined in Appendix B.
Program D — Fixed-point monitoring with individual-component survey and visual
inspection conducted regularly as outlined in Appendix B;
C. MAINTENANCE
When leaks are located by the monitoring methods described in this section, the
leaking component must then be repaired or replaced. Many components can be
serviced on-line, which is generally regarded as routine maintenance to keep
operating equipment functioning properly. Equipment failure, as indicated by a
leak not eliminated by servicing, requires isolation of the faulty equipment
for either repair or replacement.
1. Pumps
Most critical-service process pumps in SOCMI are backed up with a spare in place
so that they can be isolated for repair. Of those pumps that are not backed up
with spares, most can be isolated without major process disruption. Packed-seal
leaks frequently can be corrected by just tightening the packing, whereas mechanical-
seal leaks and packed-seal leaks that need correction require that the pump be
removed from service for seal repair. When the seal leak is small, there can be
situations in which removing the pump from service can result in larger temporary
emissions than the emissions that would occur if the pump remained in service
until shut down for other process reasons. The maintenance lead time required
to schedule repair of a leaking pump can be an important factor in the VOC emission
reduction efficiency obtainable.
2. Valves
Most valve leaks can probably be corrected on-line by tightening the packing
gland for valves with packed seals or by lubrication for plug valves, for example.
Based on field observation in one refinery study10 it was assumed that 75% of
leaking valves could be repaired on-line. Leaking valves that need to be repacked
or to be replaced for repair must be capable of being isolated or the unit must
be completely or partly shut down. Control valves, 6% of the total valves in
SOCMI, can usually be isolated.2 Block valves, which are used to isolate or
by-pass equipment, normally cannot be isolated. One refiner estimates that 10%
of the block valves can be isolated.2 Based on the assumption that there is a
-------
III-ll
random distribution of leaks versus valve type and usage, in about 3% of the
leaks the valve can be isolated and in 22% the unit will have to be shut down
for repair of the valve.
When the leaking valves can be corrected on-line, repair servicing is usually
done soon after the leak is detected. For leaks that can be corrected by isola-
tion of the valve the maintenance lead time required to schedule repair can be
an important factor in the VOC emission reduction efficiency obtainable. When
the leaks can be corrected only by a total or partial unit shutdown, the temporary
emissions could very likely be larger than the continuous emissions that would
result from not shutting down the unit until it was time for a shutdown for
other reasons.
3. Flanges
One refinery field study noted that flange leaks could be sealed effectively
on-line by simply tightening the flange bolts.10 For a flange leak that requires
off-line gasket seal replacement a total or partial shutdown of the unit would
probably be required because most flanges cannot be isolated. For many of these
cases there are temporary flange repair methods that can be used. Unless a
leak is major and cannot be temporarily corrected, the temporary emission from
shutting down a unit would probably be larger than the continuous emissions
that would result from not shutting down the unit until time for a shutdown for
other reasons. Flange leak incidences are very low and most can be corrected
by on-line maintenance.
4. Compressors
Compressors usually are in critical service and usually a spare is not provided.
In most cases the shutdown for repair of a leaking seal and the subsequent startup
will involve flaring the process stream until operations are stablized.2 This
can result in the temporary emissions being larger than the continuous emissions
that would occur until the unit was shut down for other reasons.
-------
111-12
D. REFERENCES
1. R. C. Weber, EPA, personal communications with D.. G. Erikson, Hydroscience,
Inc., Jan. 22, 1979 (analysis of preliminary results of EPA test program).
2. J. Johnson, Exxon Co., letter to Robert T. Walsh, EPA, July 28, 1977.
3. D. S. Kayser, "Rupture Disc Selection," Chemical Engineering Progress 68(5),
61—6 (1972). —
4. W.R. Seeman, Hydroscience, Inc., Flare Efficiency (memo on file at EPA, ESED,
Research Triangle Park, ND) (Sept. 29, 1978).
5. PEDCO Environmental, Inc., Cincinnati, Evaluation of Emissions from
Benzene-Related Petroleum Processing Operations (on file at EPA, ESED, Research
Triangle Park, NC) (October 1978).
6. EPA, Chemical and Petroleum Branch, OAOPS Guideline Series, Control of Volatile
Organic Compound Leaks from Petroleum Refinery Equipment, EPA-450/2-78-036,
OAQPS No. 1.2-111, (June 1978).
7. K. C. Hustuedt and R. C. Weber, "Detection of Volatile Organic Compound Emissions
from Equipment Leaks," paper presented at 71st Annual Air Pollution Control
Association Meeting, Houston, TX, June 25—30, 1978.
8. R. C. Weber,EPA, personal communications with D. G. Erikson, Hydroscience, Inc.,
Oct. 26, 1978 (analysis of preliminary results of EPA test program).
9. J. F. Lawson, Hyroscience, Inc., Trip Report for Visit to Union Carbide Corp.,
South Charleston, WV, Dec. 7, 1977 (data on file at EPA, ESED, Research Triangle
Park, NC).
10. Emissions from Leaking Valves, Flanges, Pump and Compressor Seals, and Other
Eguipment at Oil Refineries, Report No. LE-78-001, State of California Air Resource
Board (Apr. 24, 1978).
*When a reference number is used at the end of a paragraph or on a heading, it
usually refers to the entire paragraph or material under the heading. When,
however, an additional reference is required for only a certain portion of the
paragraph or captioned material, the earlier reference number may not apply to
that particular portion.
-------
IV-1
IV. COST ANALYSIS
This section presents estimated costs for the control of VOC emissions from
equipment leaks in the synthetic organic chemicals industry (SOCMI).
"A. INTRODUCTION
Estimates of capital and annual operating costs are presented for controlling
emissions from equipment leaks at SOCMI plants. The major sources of VOC emis-
sions that are considered in this section include pump and compressor seals,
in-line valves, open-ended valves, safety/relief valves, flanges, and cooling
towers. Control costs are developed for three model plants for each of the
four leak detection and maintenance programs described in Appendix B, as well
as in Sect. III.B. Costs are also presented for the five equipment modifications
discussed in Sect. III. They include the installation of double mechanical
seals, installation of a rupture disk before a safety relief valve, sampling by
the closed-loop method, capping open-ended valves, and incorporating a compressor-
degassing vent and safety relief vents into an existing closed-vent system.
Emissions reductions are not presently quantifiable since emission factors are
not available. Therefore recovered product credits and cost-effectiveness
measures have not been determined. An outline is presented that may be used to
determine recovery credits and cost effectiveness when new SOCMI emission factors
become available.
Plants in the SOCMI vary considerably in the size, configuration, and age of
facilities, the products produced, and the degree of control. Because of the
variations among plants, this cost analysis is based on three model plants,
where technical parameters are listed in Table IV-I. The parameters were selected
as being representative of existing small, medium, and large SOCMI plants based
on the data base compiled for this report. It is estimated that 15% of the
SOCMI plants would be considered to be large, 33% considered to be medium, and
52% considered to be small. Although model plant costs may differ, sometimes
appreciably, from the actual costs that may be incurred, they are the most useful
means of determining and comparing emission control costs. Plots (see Appendix C)
have been made of the total number of pumps and valves versus the plant's rated
capacity for selected plant sites in this data base. Shown on the pump plot
are the specific choices selected as model plants.
-------
IV-2
Table IV-1. Technical Parameters for Three Model Plants
ComDonent
Quantity
Small Model Plant Medium Model Plant Large Model Plant
Pumps
With packed seals
With single mechanical seals
With double mechanical seals
With no seals
Total
2
10
3
IS
6
43
10
1
60
20
133
31
1
185
Valves
Safety/relief
In-line (gas)
In-line (liquid)
Open-ended
Total
13
90
168
104
375
50
365
670
415
1500
157
1117
2074
1277
4625
Compressor seals
Flanges
Cooling towers
1
600
1
2
2400
1
a
7400
1
Components in VOC service only.
'Estimated from this report's data base.
-------
IV-3
B. CONTROL COSTS
Capital cost estimates represent the investment required to purchase and install
monitoring instruments for leak detection and the labor and material costs to
install each of the five equipment modifications.
1. Control Devices (Equipment Modifications)
Details of the capital cost estimates for the five equipment modifications are
shown in Appendix D. The cost factors used to compute annual operating costs
for all five equipment modifications are shown in Table IV-2. In computing the
depreciation and interest costs, a capital recovery factor of 0.163 was used
based on a 10-yr life at an interest rate of 10% per annum. Costs for property
taxes and insurance are computed at 4% of the capital costs. All operating
costs are for one-year periods, beginning with the first quarter of 1979.
Tables IV-3 through 5 show the installed capital cost per unit for each of the
five equipment modifications, the total installed capital cost when applied to
each model plant, and the annual operating costs incurred by each model plant
for each equipment modification. Costs for loss of production during installa-
tion or startup, preparation of the equipment for retrofitting, and other highly
variable costs are not included in the estimates.
a. Double Mechanical Seals — The installed cost to retrofit a single mechanical
seal or a packed seal is estimated to be $850.1_ 3 This cost consists of $560
for the seal and $290 labor for field installation. The cost for a new double
mechanical seal is estimated to be $575 when installed instead of a new single
mechanical seal. A credit of $2251 3 for a new single mechanical seal was
subtracted from the double-seal cost of $560 to yield a net double-seal cost of
$335. Labor to install the new double mechanical seal in a shop area was esti-
mated at $240. Actual costs for double mechaical seals will vary greatly, depending
on the pump shaft diameter, the material of construction of the seal, the pump
operating pressure and temperature, and the physical and chemical properties of
the material to be pumped. Auxiliary equipment, such as a cooler for control
of the temperature of the liquid used to flush the seal, may be required, and
can range in cost from $470 to $800;l also, an individualized pressurized pot
system may be required to maintain pressure on the liquid used to flush each
double seal, and can cost from $400 to $700.1,3
-------
IV-4
Table IV-2. Cost Factors Used in Computing Annual Costs for
Equipment Modifications (Control Devices)
Maintenance 0.05 X capital cost
Capital charges
Capital recovery 0.163 X capital cost
Miscellaneous (taxes, insurance, and 0.04 X capital cost
administration)
-------
Table IV-3. Cost Estimates for Installation of Control Devices in
a Small Model Plant
Control Device
Capital
Cost
per Unit
Number
of
Units
Total
Installed
Capital
Cost
Annual
Capital
Recovery
Cost
Annual
Miscellaneous
Cost
Annual
Maintenance
Cost
Total
Annual
Operating
Cost
Double mechanical seals
New
§ 575b
10
§5750
$ 900
$200
$300
$1400
Retrofitted
850b
10
8500
1400
300
400
2100
Rupture disk
610°
13
7930
1300
300
400
2000
Blinds on open-ended
valves
30d
104
3120
500
100
200
800
Closed-loop sampling
310d
26
8060
1300
300
400
2000
Closed-vent system
6530
1
6530
1100
300
300
1700
a
In VOC service only.
b
From refs. 1, 2, and 3.
c
From refs. 4 and 5
^Engineering estimate .
SFrom ref. 6.
-------
Table IV-4. Cost Estimates for Installation of Control Devices in
a Medium Model Plant
Control Device
Capital
Cost
per Unit
Number
of
Units
Total
Installed
Capital
Cost
Annual
Capital
Recovery
Cost
Annual
Miscellaneous
Cost
Annual
Maintenance
Cost
Total
Annual
Operating
Cost
Double mechanical seals
New $ 575fc
Retrofitted 850*
Rupture disk 610C
Blinds on open-ended 30^
valves
Closed-loop sampling 310^
£
Closed-vent system 6,530
43
43
50
415
104
1
$24,725
36,550
30,500
12,450
32,240
6,530
$4,000
6,000
5,000
2,000
5,300
1,100
$1,000
1,500
1,200
500
1,300
300
$1,200
1,800
1,500
600
1,600
300
$6,200
9,300
7,700
3,100
8,200
1,700
In VOC service only.
"Worn refs. 1, 2, and 3.
'From refs. 4 and 5.
^Engineering estimate.
'From ref. 6.
-------
Table IV-5. Cost Estimates for Installation of Control Devices in a Large Model Plant
Control Device
Capital
Cost
per Unit
Number
of
Units3
Total
Installed
Capital
Cost
Annual
Capital
Recovery
Cost
Annual
Miscellaneous
Cost
Annual
Maintenance
Cost
Total
Operating
Cost
Double mechanical seals
New
$ 57 5b
133
$ 76,475
$12,500
$3,100
$3,800
$19,400
Retrofitted
850b
133
113,050
18,400
4,500
5,700
28,600
Rupture disk
610C
157
95,770
15,600
3,800
4,800
24,200
Blinds on open-ended
valves
30d
1,277
38,310
6,200
1,500
1,900
9,600
Closed-loop sampling
310d
319
98,890
16,100
4,000
4,900
25,000
Closed-vent system
6,530®
1
6,530
1,100
300
300
1,700
a
In VOC service only.
^From refs. 1, 2, and 3.
c
From refs. 4 and 5.
d
Engineering estimate.
e
From ref. 6.
-------
iv-a
Rupture Disks — The cost for a rupture disk located upstream of a relief
valve is estimated to be $6104'5 for either a new or a retrofitted installation.
The disk itself with a holder costs $330. The pressure gauge and vent valve
assembly add $40, and labor to install the disk is estimated at $240. It has
been assumed that no piping modifications are required and that the disk and
its holder can simply be inserted between the flanges of the relief valve and
the system it protects. Actual costs for rupture disks can vary greatly, depending
on the operating pressure required, the number of disks purchased, the material
of construction required, and the size of the disk itself. The rupture disk
can become a cost saving installation if it allows the relief valve to be tested
in place instead of having to be removed to a shop area. This is normally done
once or twice annually.
Plugs on Open-Ended Valves — The installed cost of a blind flange on an open-ended
valve is estimated to be $30. This cost includes the blind flange and the bolt
and gasket set and is representative of either a new or retrofitted installation.
Actual costs for blinding or plugging open-ended valves depend on the device
used, i.e., plug, cap, or blind flange, the size of the valve, and the material
of construction required.
Closed-Loop Sampling — The cost of a closed-loop sampling system is estimated
to be $310 for either a new or a retrofitted installation. The costs are based
on a 6-m length of 2.5-cm-diam, schedule 40, carbon steel pipe and three 2.5-cm-diam
carbon steel ball valves. It is estimated from the data base that approximately
25% of the open-ended valves are used for sampling. Actual costs will be highly
variable, depending on pipe size and material of construction required.
Closed-Vent System — The installation of a vent line from the compressor oil-seal
reservoir and from the plant safety relief valves to an existing (retrofitted)
controlled closed-vent system is estimated to be $6530.8 This cost is based on
installing a L22-m length of 5.1-cm-diam, schedule 40, carbon steel pipe at a
cost of $5200 and three 5.1-cm-diam cast steel plug valves and one metal gauze
flame arrestor 5.1 cm in diameter at a cost of $1330. It is estimated that
only one vent line with a header system is necessary for collecting the vapors.
Actual costs will depend on the line size required, the locations of relief
-------
IV-9
valves and compressor reservoirs, and the materials of construction. If an exist-
ing controlled closed-vent system is not available, the installation costs will be
significantly higher. As an example, a new flare system for a model 600-Gg/yr
ethylbenzene/styrene plant was estimated to cost $272,0007 and one for a 226.8-Gg/yr
ethylene plant to cost $145,000.®
2. Leak Detection and Maintenance Programs
The cost factors used to compute annual operating cost for all four leak detec-
tion programs are shown in Table IV-6. These factors are based on an EPA report6
and on engineering estimates. In computing the depreciation and interest costs
a capital recovery factor of 0.163 was used based on a 10-year life at an interest
rate of 10% per annum for the fixed-point monitoring system. A Capital recovery
factor of 0.23 was used, based on a 6-yr life at an interest rate of 10% per
annum, for the portable VOC analyzers. Costs for property taxes and insurance
are computed at 4% of the capital costs. All operating costs are for 1-year
periods, beginning with the first quarter of 1979. Tables IV-7 through IV-10
show the installed capital cost for each of the four monitoring programs, the
annual instrument capital recovery costs, the annual instrument materials, main-
tenance and calibration costs, the annual monitoring labor costs, the annual
maintenance labor cost, the annual administrative and support costs, and the
total annual operating costs for each of the three model plants.
a. Individual-Component Survey with Visual Inspection (Program A) — The installed
capital cost for instruments shown in Table IV-7 is $9500, based on the use of
two portable organic analyzers (one used as a spare) at a cost of $4750 each.6
The annual cost for instrument materials, maintenance, and calibration is $2700,6
including replacement of one battery pack and two filter packs per year. The
annual monitoring labor cost for each model plant was obtained by multiplying
the total monitoring manpower requirements (from Table B-2, Appendix B) by $15/hr.6
The annual maintenance labor cost for each model plant was obtained by multiplying
the total maintenance manpower requirements (from Table B-7, Appendix B) by
$l5/hr.6 (Refer to Appendix B for description of monitoring method.)
b. Fixed-Point Monitoring System with Individual-Component Survey Conducted Initially
(Program B) — The installed instrument capital cost shown in Table IV-8 is
$137,000, which includes $9,500 for two portable VOC analyzers and $127,5009
-------
IV-10
Table IV-6. Factors Used in Computing Annual Costs
for Monitoring Programs
Monitoring instrument capital
Portable
Fixed point (installed)
Labor
Monitoring
Maintenance
Annual instrument capital-related cost'
Capital recovery
Portable monitor
Fixed-point monitor
Miscellaneous (taxes and insurance)
Annual instrument material
maintenance and calibrations
Portable monitor
Fixed-point monitor
Annual administrative and support
$4750
$127,500*
$15.00/hr
$lS.00/hr3
0.23 X capital
0.168 X capital
0.04 X capital
$2700
0.05 X capital cost
0.4 X annual monitoring
and maintenance cost
From ref. 6.
3From ref. 9.
-------
IV-11
Table IV-7. Control Cost Estimates of Monitoring and Maintenance
Program A for Three Model Plants
Program A: Individual-Component Survey.with Visual Inspection
Cost
Small
Model Plant
Medium
Model Plant
Large
Model Plant
Instrument capital cost
$9,500
$ 9,500
$ 9,500
Annual instrument capital-related cost
$2,600
$ 2,600
$ 2,600
Annual instrument materials maintenance
and calibration cost
2,700
2,700
2,700
Annual monitoring labor cost
690
2,650
8,250
Annual maintenance labor cost
2,550
10,200
34,500
Annual administrative and support cost
1,300
5,150
17,100
Total annual cost
$9,840
$23,300
$65,150
-------
IV-12
Table IV-3. Control Cost Estimates of Monitoring and Maintenance Program B for
Three Model Plants
Program B: A Fixed-Point Monitoring System with an Individual-Component
Survey Conducted Initially
Instrument capital
Annual instrument capital-related cost
Annual instrument material
maintenance and calibration
Cost
Small Medium Large
Model Plant Model Plant Model Plant
5137,000 $137,000 $137,000
$28,500 $28,500 $28,500
9,100 9,100 * 9,100
750
11,385
4,850
$54,555
$26,490
Annual monitor labor*
Annual maintenance labor*
Annual administrative and support
Total annual costs
One-time maintenance/monitoring
labor costs for initial ICS
(1st year only)
60
340
360
$38,360
$ 1,995
240
3,365
1,440
$42,645
$ 7,915
~Assumes that 33% of the leaks are detected.
-------
IV-13
Table IV-9. Control Cost Estimates of Monitoring and Maintanance Program C for
Three Model Plants
Program C: Unit Area Survey with Individual-Component Survey Conducted Initially
Cost
Small
Model Plant
Medium
Model Plant
Large
Model Plant
Instrument capital
$9,500
$9,500
$ 9,500
Annual instrument capital-related cost
$2,600
$2,600
$ 2,600
Annual instrument materials
maintenance and calibration
2,700
2,700
2,700
Annual monitor labor*
570
1,665
4,665
Annual maintenance labor*
1,275
5,100
17,250
Annual administrative and support
740
2,700
8,800
Total annual costs
$7,885
$14,765
$36,015
One-time monitoring/maintenance
labor cost (1st year)
$1,560
$ 6,180
$20,625
~Assumes that 50% of the leaks are detected.
-------
IV-14
Table IV-10. Control Cost Estimates of Monitoring and
Maintenance Program D for Three Model Plants
Program D: Fixed-Point. Monitoring System with Individual
Component Survey and Visual Inspection Conducted Regularly*
Cost
Small
Plant
Medium
Plant
Large
Plant
Instrument capital
$137,000
$137,000
$137,000
Annual instrument capital-related cost
$28,500
$28,500
$28,500
Annual instrument materials maintenance
and calibration
9,100
9,100
9,100
Annual monitor labor
750
2,895
9,000
Annual maintenance labor
2,550
10,200
34,500
Annual administrative and support
1,320
5,240
17,400
Total annual costs
$42,220
$55,935
$98,500
~Assumes that fixed-point monitor detects 33% of the leaks and that ICS and
visual detects the remainder»
-------
IV-15
for a 20-point gas chromatograph fixed-point monitor. The annual instrument
capital recovery cost of $28,500 consists of $2,600 for the portable analyzers
and $25,900 for the fixed-point unit. Annual instrument materials, maintenance,
and calibration costs of $9100 reflect a cost of $2700 for the portable analyzer
and $6400 for the fixed-point monitor.
As discussed Sect. III.B, the fixed-point monitor will detect 33% of the signifi-
cant leaks. The labor costs for monitoring and maintenance reflect the costs
expected during the second year of operation with 33% of the leaks detected.
The monitoring labor costs were obtained by multiplying the total annual monitor-
ing requirement for the FPM only from (Table B-3) by $15/hr. The maintenance
labor cost was obtained by multiplying the annual maintenance lafior requirement
(from Table B-7) by 33% and by $15/hr. For the first year only, the individual-
component survey monitoring labor (Table B-3) and maintenance labor (Table B-7)
costs will be incurred. (Refer to Appendix B for monitoring method.)
Unit Area Survey with Individual-Component Survey Conducted Initially (Program C) —
The installed instrument capital cost of $9500 shown in Table IV-9 is for the
two portable analyzers discussed in Program A. As was discussed in Sect. III-B,
the unit area walk-through survey will detect 50% of the significant leaks.
The costs for monitoring labor and maintenance labor reflect the costs expected
during the second year of operation. With 50% of the leaks detected the monitor-
ing labor costs were obtained by multiplying the total annual monitoring manpower
requirements for the unit area survey only (from Table B-5) by $15/hr. The
maintenance labor cost was obtained by multiplying the annual maintenance manpower
requirement (from Table B-7) by 50% and by $15/hr. For the first year only,
the individual-component survey monitoring labor and maintenance labor costs
(Table B-7) will be incurred. (Refer to Appendix B for monitoring method.)
Fixed-Point Monitoring System with Individual-Component Survey and Visual
Inspection Conducted Regularly (Program D) — The installed capital cost of
$137,000 shown in Table IV-10 of $137,000 is identical to that discussed in
Program B. The annual monitoring labor costs for each model plant were obtained
by multiplying the total monitoring manpower requirement (from Table B-6) by
$15/hr. The annual maintenance labor cost was obtained by multiplying the total
maintenance manpower requirements (from Table B-7) by $15/hr. (Refer to Appendix B
for description of monitoring method.)
-------
IV-16
The first-year operating costs for each of the monitoring programs discussed
will be higher than were estimated if the number of initial leaks is higher
than the number of leaks that recur after the monitoring programs are initiated.
C. CONTROL IMPACT
The environmental and cost impacts for the controls discussed in this report
cannot be assessed because at present adequate emission factors are not available.
Field studies are in progress to provide these factors.10 Factors from refinery
studies in 1957—1958 probably are not representative of industry today.11 The
following is a discussion of the major factors that determine the impacts.
1. Equipment Modification Controls
Equipment modification controls apply only to specific equipment components or
procedures in the model plants. For double mechanical seals on pumps, rupture
disks upstream of relief valves, and blinds on open-ended valves, the leaks are
essentially eliminated and the control efficiency is high, probably approaching
100%. For a closed-vent system the emissions from the leaks are destroyed, and
the efficiency depends on the ultimate control device. For a flare as the ulti-
mate control device, this is probably 98 to 99%. For closed-loop sampling the
efficiency should be high. The annual emission reduction, then, is mainly a
function of the leaks affected by the modification. The annual operating cost
for each equipment modification for the model plants is discussed in Sect. IV-B.
The cost effectiveness is the ratio of these costs, adjusted for recovery credit,
to the annual emission reductions achieved by the control.
2. Leak Detection and Maintenance Controls
The annual emission reduction for the model plants is a function of the leak
rate of the repaired component, the number of leaks corrected, and the time
that the leaks would have been active if not corrected. For all control methods
a leak is designated to be corrected if the measured VOC concentration at the
source is <10,000 ppm. Analysis of preliminary field study data indicates the
emission rates to be 1.11 kg/day for pump seals and 0.19 kg/day for valves at a
10,000-ppm concentration leak.® The average emission rate for all leaks
>10,000 ppm will be higher, possibly by 1 order of magnitude. The number of
leaks corrected depends on the occurrence frequency and the detection efficiency.
The occurrence frequency is discussed in Table B-7 and the detection efficiency
-------
IV-17
depends on the detection method used, as is discussed in Sect. III-B. If the
leak occurrences are at a constant rate, the maximum annual average time that a
detected leak would be active if not corrected would be 6 months. From this
average time must be deducted the average time that the leak is active before
it is detected, which is dependent on the leak detection procedure used, and
the average time that the leak is active after it is detected, which is dependent
on the maintenance scheduled used. The maintenance schedule used must, as is
discussed generally in Sect. III-C, take into consideration that the potential
temporary emissions could be excessive in relation to the continuous leak emissions.
To obtain the emission control efficiency the emissions reduced are compared to
those that would have occurred if there had been no control. The maximum control
efficiency of leak detection and maintenance controls discussed depends on the
minimum size of the leak designated to be repaired (>10,000 ppm at the source).
Recent refinery tests indicate that only a small percentage of the leaks account
for about 90% of the total VOC emissions.6 Because the control methods discussed
are based on locating the large leaks, the maximum control efficiency attainable
is probably about 90%. The efficiency for each control method discussed will
be less, depending on the time lapse before the leaks are detected, the efficiency
in locating the leaks, and the time it takes to correct the leaks after detection.
The cost-effectiveness ratio is the net total annual' cost divided by the annual
emission reduction. The annual operating cost for each leak detection and main-
tenance control method for the model plants is discussed in Sect. IV-B. The
net total annual cost is the residual after credit is taken for the material
saved. The 1978 average value of products in SOCMI is $331/mg,- this amount
times the annual emission reduction gives the control method credit.
-------
17-18
D. REFERENCES
1. E. MacRae, Durametallic Corp., Kalamazoo, MI, personal communcation with D. G.
Erikson, Hydroscience, Inc., Jan. 17, 1979 (documented for files of D. G. Erikson).
2. D. Way, Crane Corp., Morton Grove, IL, personal communication with D. G. Erikson,
Hydroscience, Inc., Jan. 16, 1979 (documented for files of D. G. Erikson).
3. M. Petin, Johnson City, TN, representative of Chesterton Co., personal communica-
tion with D. G. Erikson, Hydroscience, Inc., Jan. 23, 1979 (documented for files
of D. G. Erikson).
4. J. Smith, Atlanta, GA, representative of Continental Disc Corp., personal communi-
cations with D. G. Erikson, Hydroscience, Inc., Jan. 19, 1979 (documented for
files of D. G. Erikson).
5. E. Wischhusen, Knoxville, TN, representative of 8S&B Safety Systems, Inc., per-
sonal communication with D. G. Erikson, Hydroscience, Inc., Jan. 22, 1979 (docu-
mented for files of D. G. Erikson).
6. EPA, Chemical and Petroleum Branch, OAQPS Guideline Series. Control of Volatile
Organic Compound Leaks from Petroleum Refinery Equipment, EPA-450/2-78-036,
OAQPS No. 1.2-111 (June 1978).
7. J. A. Key, Hydroscience, Inc., Emission Control Options for the Synthetic
Organic Chemicals Industry. Ethylene Benzene/Styrene Product Report (on file
at EPA, ESED, Research Triangle Park, NC) (1978).
8. R. L. Standifer, Hydroscience, Inc., Emission Control Options for the Synthetic
Organic Chemicals Industry. Ethylene Product Report (on file at EPA, ESED,
Research Triangle Park, NC) (1978).
9. K. C. Hustuedt and R. C. Weber, Detection of Volatile Organic Compound Emissions
from Equipment Leaks, paper presented at 71st Annual Air Pollution Control
Association Meeting, Houston, TX, June 25—30, 1978.
10. Radian Corp., Assessment of Environmental Emissions from Oil Refining,
EPA Contract No. 68-02-2665, in progress, March 1976 to March 1979.
11. Los Angeles County Air Pollution Control District, Joint District, Federal and
State Project for the Evaluation of Refinery Emissions (nine reports) (1957—1958).
*When a reference number is used at the end of a paragraph or on a heading,
it usually refers to the entire paragraph or material under the heading.
When, however, an additional reference is required for only a certain portion
of the paragraph or captioned material, the earlier reference number may not
apply to that particular portion.
-------
V-l
V. ASSESSMENT
A. SUMMARY
Fugitive emissions are VOC emissions that result from leaks from plant equipment.
Equipment characterized in this report include pump seals, in-line valves, open-
ended valves, safety/relief valves, compressor seals, flanges, and cooling towers.
The total estimate of equipment components in VOC service in the SOCMI includes
46,690 pumps, 1,154,000 valves, 1,846,200 flanges, 1,870 compressors, and 1,334
cooling towers.
Control of fugitive VOC emissions is primarily achieved by use of leak detection
and maintenance programs. They range from complete testing of every potential
VOC leak source to use of fixed-point monitors installed to detect leaks in
specific process unit areas. Leak detection efficiencies range from 100 to
33%. A significant leak is defined as >10,000 ppra read out on a portable VOC
analyzer at the source of the leak (0 cm distance). Since emission factors are
not presently quantifiable, emission reduction efficiencies could not be estimated.
In addition to leak monitoring and maintenance, five specific equipment modifi-
cations can be used as a supplement to eliminate VOC emissions specific to those
sources.
The costs for control of fugitive emissions from the three model SOCMI plants
range from $800 to $28,600 annually for the equipment modifications and from
$7,800 to $98,500 for the various leak detection methods. Installed capital
costs range from $3,100 to $113,000 for the equipment modifications and from
$9,500 to $137,000 for the leak detection programs.
B. -SUPPLEMENTAL INFORMATION
Fugitive VOC emissions from SOCMI plants are currently being measured as part
of an emission assessment program that Monsanto Research Corp.1 is conducting
under contract with the U.S. Environmental Protection Agency. In addition.
Radian Corp.2 is conducting a similar, yet more extensive, testing program for
refineries also under contract to the EPA. Results of both programs should
provide data on fugitive emissions that reflect current technology and operating
practices. Emission factors will also be developed from the Radian data which,
-------
V-2
when applied to the industry characterization in Sect. II of this report, should
give a representative estimate of current total VOC emissions from SOCMI. They
should also provide a basis for estimating maintenance manpower requirements by
supplying current leak occurrence frequencies for each equipment source category
in the SOCMI and enable the effectiveness of control factors to be quantified.
-------
V-3
C. REFERENCES*
1. Monsanto Research Corp., Measurement of Fugitive Emissions from Petrochemical
Plants, EPA Contract No. 68-02-1874, in progress.
2. Radian Corp., Assessment of Environmental Emissions from Oil Refining, EPA
Contract No. 68-02-2665, in progress, March 1976 to March 1979.
*When a reference number is used at the end of a paragraph or on a heading,
it usually refers to the entire paragraph or material under the heading.
When, however, an additional reference is required for only a certain portion
of the paragraph or captioned material, the earlier reference number may not
apply to that particular portion.
-------
APPENDIX A
TOTAL NUMBER OF SOCMI PRODUCT SITE LOCATIONS
-------
A-3
Table A-l. Total Number of SOCMI Product Site Locations3
Location
Number of Product Sites
Texas
216
New Jersey
136
Louisiana
94
Illinois
90
Ohio
76
California
74
Pennsylvania
71
New York
64
West Virginia
37
Tennessee
32
Alabama
32
North Carolina
32
Kentucky
30
Indiana
28
Michigan
25
Puerto Rico
23
All others
271
Total
1334
Radian Corp., Austin, TX, Organic Chemical Producers'
Data Base Program, Vol. II, Final Report (data on file
at EPA, ESED, Research Triangle Park, NC) (Aug. 5, 1976).
^A product site is defined as a geographic location
that requires specific pieces of equipment to produce
a specific synthetic organic chemical.
-------
APPENDIX B
MONITORING/MAINTENANCE MANPOWER REQUIREMENTS
FOR FOUR MONITORING PROGRAMS
-------
B-3
APPENDIX B
A. MONITORING REQUIREMENTS
The manpower required to monitor equipment leaks varies with the methods or
combination of methods used. The manpower required for the four monitoring
programs outlined in Sect. III.B have been estimated for each of the three model
plants discussed in Sect. IV. These estimates are based on an EPA report,1 EPA
estimates,2 and the monitoring guidelines discussed in Sect. III. Table B-l
shows the maximum number of leaks measuring >10,000 ppm at the leak surface
available for detection per year for each of the three model plants.1 The total
number of leaks expected annually for each model plant has a major impact on
the amount of time required for monitoring.
1. Program A — Individual Component Survey with Visual Inspection
Table B-2 shows the annual monitoring manpower requirements (man-hours) required
to perform an individual-component survey with a portable organic analyzer on
each of the three model plants. Two analyzers are required, one for use and
one as a spare. Pump seals, in addition to being checked once a year with the
instrument, are visually inspected for leakage once each week. Compressor seals,
in-line valves in gas service, and safety relief valves are checked quarterly
with the instrument. Flanges are not monitored since the leak occurrence is
considered negligible. For the purposes of these estimates only, it is assumed
that this survey will be conducted by two people — one operating the portable
VOC detection instrument and the other recording the results. The visual inspec-
tions are assumed to be performed by one person.
2. Program B — Fixed-Point Monitoring System with an Individual-Component Survey
Conducted Initially
Table B-3 shows the annual monitoring manpower requirements for a fixed-point
monitoring system with an individual-component survey (ICS) conducted initially.
It is assumed that the ICS will locate all significant leaks and reduce the VOC
background to allow for reliable detection of recurring leaks by the fixed-point
continuous monitor. The fixed-point monitor in use is assumed to have 20 points
strategically located throughout the process unit to give the best allowable
coverage in areas where, historically, leakage has been a problem. The man-hours
required for monitoring after the first year are based on the assumption that the
-------
B-4
Table B-l. Basis for Determining Monitoring/Maintenance
Manpower Requirements
Components
Number of
ComDonentsa
Leakage
Frequency
-------
Table B-2. Annual Monitoring Manpower Requirements for Individual-Component Survey with
Visual Inspection (Program A) for Three Model Plants
Components
No.
per
of Components
Model Plant3
Type of
Monitoring
Estimated Time^
for Monitoring
(min)
Times
Monitored
(No./yr)b
Annual
(hr)
Manpower
per Model
Required
Plantc
Small
Medium
Large
Small
Medium
Large
Pump seals
15
60
185
Instrument
5
1
2.5
10
30.8
Visual
0.5
5?
6.5d
26d
00.2'
Compressor seals
1
2
8
Instrument
10
4
1.3
3
10.7
In-line valves (gas)
90
365
1117
Instrument
1
4
12
49
148.9
In-line valves (liquid)
168
670
2074
Instrument
1
1
5.6
22
69.1
Open-ended valves
104
415
1277
Instrument.
1
1
3.5
14
42.6
Safety/relief valves
13
50
157
Instrument
8
4
13.9
53
167.5
Pipeline flanges
600
2400
7400
None
Cooling tower
1
1
1
Instrument
10
1
0.3
0.3
0.3
Total
46
177
550
a
From report data base.
bSee ref. 1.
Q
Except where otherwise noted, based on 2 persons required.
d
Only one person required.
-------
Table B-3. Annual Monitoring Manpower Requirements for a Fixed-Point Monitoring (FPM) System with an
Individual-Component Survey Conducted Initially {Program B)
Ho.
per
of Components
Model Planta
Estimated Time^ Times
Type of for Monitoring Monitored
Monitoring (itiin) (No./yr)'3
Annual Manpower Required
(hr) per Model Plantc
Components
Small
Medium
Large
Small
Medium
Large
Individual-Component Survey (ICS) Conducted Initially
Pump seals
15
60
185
Instrument 5 1
2.5
10
30.a
Compressor seals
1
2
a
Instrument 10 1
0.3
0.7
2.7
In-line valves (gas)
90
365
1117
Instrument 1 1
3
12
37.2
In-line valves (liquid)
168
670
2074
Instrument 1 1
5.6
22
69.1
Open-ended valves
104
415
1277
Instrument 1 1
3.5
14
42.6
Safety/relief valves
13
50
157
Instrument 8 1
3.5
13
41.9
Pipeline flanges
600
2400
7400
None
Cooling tower
1
I
1
Instrument 10 1
0.3
0.3
0. 3
Tofal ICS
Fixed-Point¦Monitoring
19
72
225
Leaks Detected (Mo./yr)
per Model Plantd
Tirae Required to Locate
Each Leak (hr)e
Annual
(he)
Manpower
per Model
Required
Plantc
Activity
Small
Medium
Large
Small
Medium
Large
Locating leaks detected
a
32
100
0.5
4
16
50
Total FPM and ICS
combined
(1st year only)
23
aa
275
From report data base. bSee ref. 1. CTwo persons required. dAssumes 33%. detection efficiency. eSee ref. 2.
-------
B-7
fixed-point monitor will detect 33% of the leaks annually.3 The man-hours estimated
for the individual-component survey occur in the first year of operation only. The
hours required are based on an assumption that 1/2 hr is spent in identifying the
leaking source once the fixed-point monitor alarms. Specific leak detection is
done with two portable monitors (one is a spare).
3. Program C — Unit Area Survey with Individual-Component Survey Conducted Initially
Table B-4 shows the time required to set up, conduct, and complete a unit area
survey.2 The development of a walk-through path is considered to be a one-time-only
effort requiring 2 hr by one man, regardless of the size of the process unit.
The unit area walk-through survey itself is conducted with a portable organic
analyzer. Two analyzers are required, one for use and one as a Spare. The
survey is conducted twice a month back to back to minimize the chance of signifi-
cant leaks going undetected. The time required for the actual walk-through
survey is estimated to be 115 min per man (2 men required) per pump in the process
unit.2 The survey results would need to be analyzed for each walk-through, which
would require 1/2 hr each. Locating the leaks is estimated to require 1/2 hr
for each leak (peak) on the monitor printout and would be done only once each
month after back-to-back walk-throughs were done. Leak detection efficiency
for a unit area survey is estimated to be 50% of all significant leaks.2
Table B-5 shows the annual monitoring manpower required for a unit area survey
with an individual-component survey conducted initially. Thus the total monitoring
man-hours for all three model plants is the time required for the first year of
operation. From the second year on, only the unit area survey monitoring would
be required, based on detecting only 50% of the recurring leaks annually.
4. Program D — Fixed-Point Monitoring System with an Individual-Component Survey
and Visual Inspection Conducted Regularly
Table B-6 shows the annual monitoring manpower requirements for a fixed-point
monitor with selected quarterly individual-component surveys by instrument and
weekly visual inspections. This program is the most labor-intensive of the
four programs because it includes Program A every year plus the continuous fixed-
point monitor described in Program B. Conversely, it should have the best control
effectiveness of the four programs described because it not only detects 100% of
all annual recurring leaks but also detects 33% of them as they occur, thus
minimizing the time that the leak exists before it is corrected.
-------
B-3
Table B-4. Monitoring Manpower Requirements for
One Unit Area Survey*
Time Required
2.0 hr
3.0 min
0.5 hr
0.5 hr
peaks (per peak1)
Development of walk-through path
(nonrecurring, one time only)
Walk-through survey (per pump)
Analysis of survey results (one man)
Locating leaks identified by chart
*See ref. 2.
-------
B-9
Table B-5. Annual Monitoring Manpower Requirements for a Unit Area Survey (UAS)
with an Individual-Component Survey (ICS) Conducted Initially (Program C)
Components
No. of Components
per Model Plant3
Small Medium Large
Type of
Monitoring
Estimated Time^
for Monitoring
(min)
Times
Monitored
(No./yr)b Small
Annual Manpower Required
(hr) per Model Plant0
Medium
Larae
Individual-Component Survey (ICS) Conducted Initially
Pump seals
IS
60
185
Instrument
5
1
2.5
10
30.8
Compressor seals
1
2
8
Instrument
10
1
0.3
0.7
2.7
'in-line valves (gas)
90
365
1117
Instrument
1
1
3
12
37.2
In-line valves (liquid)
168
670
2074
Instrument
1
1
5.6
22
69.1
Open-ended valves
104
415
1277
Instrument
1
1
3.5
14
42.6
Safety/relief valves
13
50
157
Instrument
8
1
3.5
13
41.9
Pipeline flanges
600
2400
7400
None
Cooling tower
1
1
1
Instrument
10
1
%
0.3
0.3
0.3
Total ICS
19 ,
72
225
Activity
Number of Tines
Activity Performed
Per Year
d
Unit Area Survey
Time Required for Each
Activity (hr) per Model Plant
Annual
(hr)
Manpower
per Model
Required
Plantc
small Medium Large
Small
Medium
Large
Development of walk-
1
2 2 2
2
2
2
through path
Walk-through survey
24
0.75 3 9.25
18
72
222
Analysis of survey results
24
0.5 0.5 0.5
12
12
12
Locating leaks identified
12
0.5 2.1 6.25
6
25
75
by chart peaks (50*
detected)
Total OAS (from 2nd year
on)
38
111
311
Total UAS and ICS combined (1st year only)
57
183
536
*From report data base.
bSee ref. 1.
cTVfo persons required.
dSee ref. 2.
-------
Table B-6. Annual Monitoring Manpower Requirements for a Fixed-Point Monitoring (FPM) System
with an Individual-Component Survey (ICS) and Visual Inspection Conducted Regularly (Program D>
Components
No. of Components
par Model Planta
Small Medium Large
Type of
Moni torlng
Estimated Time,
£
for Monitoring
(min)
Times
Monitored
(No./yr)b
hnnual
(hr)
Manpower Required
per Model Plant0
Small
M ed i urn
Large
2.5
6.5d
10
26d
30.8
80. 2d
L,3
3
10.7
12
49
148.9
5.6
22
69.1
3.5
14
42.6
13.9
53
167.5
0-3
0.3
0. 3
46
177
550
Individual-Component Survey with Visual Inspection
Pump seals
15
60
185
Instrument
5
visual
0
Compressor seals
1
2
8
Instrument
10
In-line valves (gas)
90
365
1117
Instrument
1
In-line valves (liquid)
168
670
2074
Instrument
1
Open-ended valves
104
415
1277
Instrument
1
Safety/relief valves
13
50
157
Instrument
e
Pipeline flanges
600
2400
7400
None
Cooling tower
1
1
1
Instrument
10
1
52
4
4
1
1
4
td
I
J-1
O
lota I
Fixed-Point Monitoring
Activity
Locating Leaks detected
Leaks Detected per Year by the
Fixed-Point Monitor per Model Plants4
Small Medium Large
8 32
Total FPM and ICS with visual inspection regularly
100
Time Required to
Locate Each Leak (hr)
0.5
From report data base. See ref. 1.
Annual Manpower Required
(hr) per Model Plant0
Small
Medium
Large
_4_
50
16
193
50
600
"Two persons required unless otherwise noted.
Assumes that FPM detects 33i of leaks and that ICS and visual detects the remainder.
Only one person required.
See ref. 2.
-------
B-ll
B. MAINTENANCE REQUIREMENTS
It should be noted that no matter what monitoring program is used associated
maintenance manpower will be required for repairing the leaking component.
These requirements (shown in Table B-7) are estimated from an EPA report1 and
the assumptions noted in the table. The actual maintenance manpower required
for each program is directly proportional to the percentage of leaks detected;
i.e, if only one-half the significant leaks are detected, then the maintenance
man-hours will be 50% of that shown in Table B-7.
-------
Table B-7. Annual Maintenance Manpower Requirements for
Three Model Plants
Components
per Model
Plant
Leakage
Repair
Annual
Manpower Requirement
(hr)°
Small
Medium
Large
Frequency
Time
Small
Medium
Large
Component
Plant
Plant
Plant
(%/yr.)a
(hr/leak)
Plant
P] ant
Plant
Pump seals
15
60
185
12
80
145
575
17 BO
Process valves
362
1450
4468
6
1.13d
25
100
300
Safety/relief valves
13
50
157
7
oe
Flanges
600
2400
7400
Minor
Compressor seals
1
2
8
7
40
5
20
Cooling towers
1
1
1
f
Total
170
680
2300
aBased on the number of leaks (>10,000 ppm at the source) detected in refinery tests (ref. 1) and the assumption
that the leak recurrence is annual at this percentage of component leakage.
b . .
See ref. 1.
Q
Based on all leaks being detected and corrected.
Weighted average based on 75% of the leaks repaired.on-line, requiring 0.17 man-hour per repair, and on 25% of the
leaks repaired off-line, requiring 4 man-hours per repair.
Q
It is assumed that these leaks are corrected by routine maintenance at no additional man-hour requirements.
^Leaks are detected only at the cooling tower but occur somewhere in the cooling water system; no data on occurrence.
-------
B-13
C. REFERENCES*
1. EPA, Chemical and Petroleum Branch, OAQPS Guideline Series. Control of Volatile
Organic Compound Leaks from Petroleum Refinery Equipment, EPA-450/2-78-036, OAQPS
No. 1.2-111 (June 1978).
2. R. C Weber, EPA, personal communications with D. G. Erikson, Hydroscience, Inc.,
Oct. 26, 1978 (analysis of preliminry results of EPA test program).
3. K. C. Hustuedt and R. C. Weber, "Detection of Volatile Organic Compound Emissions
from Equipment Leaks," paper presented at 71st Annual Air Pollution Control
Association Meeting, Houston, TX, June 25—30, 1978.
*When a reference number is used at the end. of a paragraph or on a heading,
it usually refers to the entire paragraph or material under the heading.
When, however, an additional reference is required for only a certain portion
of the paragraph or captioned material, the earlier reference number may not
apply to that particular portion.
-------
APPENDIX C
NUMBER OF PUMPS AND VALVES VERSUS
PLANT CAPACITY AND TYPICAL MODEL PLANTS
-------
C-3
a
Table C-l. Model Plants Characterized to Date
Total Number
b
of Components
Plant
Plant Type
Pumps
Process
Valves
Relief
Valves
Compressors
Capacity
(Gg/yr)
Acrylic acid and esters
36
820
44
76.1
Acrylonitrile
50
1200
180
Adipic acid
56
349
22
150
Chlorobenzenes
102
800
12
96
Chloromethanes
30
1000
15
2
45—180
Cyclohexane-benzene
15
300
15
1
30—265
Cyclohexane-petroleum
35
700
35
1
100
Cyclohexanol/cyclo-
hexanone — cvclohexane
75
1975
56
100
Cyclohexanol/cyclo-
hexanone — phenol
35
875
26
100
Ethylbenzene/styrene
50
1000
300
Ethylene
165
4150
65
8
408.2—680.
Ethylene dichloride
42
1100
40
400
Ethylene oxide — air
10
400
2
227
Ethylene oxide — oxygen
10
400
2
136.1
Formaldehyde
13
214
6
45
Maleic anhydride
15
500
22.7
Nitrobenzene
42
500
20
30—150
a
From Emissions Control Options
for the Synthetic Organic
Chemicals Manufacturing
Industry — Product Reports (on file at EPA, ESED, Research Triangle Park, NC) .
In VOC service only.
-------
C-4
Log Y » .82 + .27 Log X (Slope)
r 13 0.07 (Correlation Coefficient)
lOOO
500
- IOO
50
IO
•
•
. large.
PUAKJT
5 £
* 8
LL-
*
1
<
>
<
i
•
>
•
•
m
- P-1C.LJI UIV1 MUU&U
m
•
m
• •
m
PUAWT 53 <7.
•
•
f
•
<
# '
L
1
>•
i
•
+
•
-¦ <
~
m
SMALL
PLAKiT
MOC
5Z •
>ej-
o
#
4
4
~-
l
i
~
•-
»
•
<
1
_
lO
50 IOO
Plant Capacity M lbs (X)
5 CO
lOOO
SOC
Fig. C-l. Total Number of Pumps for Each Plant in the Data Base Versus
the Plant's Rated Capacity (54 Points!
-------
C-5
¦u
c
(0
u
03
CU
in
»H
(0
>
O
o
z
•
•
•
•
~
¦
>
t
•
*
m
f
•
•
•
0
i 4
\
•
*
i
f
€
•
•
«
>
•
W
#'
m
-
•
m
•
i
-
lO
too
Plant Capacity M lbs (X)
\ooo
:;50<
Fig. C-2. Total Number of Valves for Each Plant in the Data Base Versus
the Plant's Rated Capacity (42 Points)
-------
APPENDIX D
COST ESTIMATE DETAILS
-------
PRELIMINARY CAPITAL
D-3
vm—
of
"Oo^jJolg. w\«e.g.V>o.»\' c-g-^ ^go-\ ^ - Mguo »nJlS<.4-/ofi 4—
PHASE
e f NUM»e«
ST
OaTE
JC3 NU«6tfl
J A-*. "2.Z. («J"l °l
9 ! CI- £¦
CaSE NU»6ER
SECTION NUMBER
SECTOR NUMBER 4 nau£
• 11
NAME OF F ACILITY! D E SCR I P Tl 0 N/0UANf.|
»vle. el>v>. c_«>-i ^
\^H *U A.+--V- Ce/a.'Vi.CVs c^.wl,«rs ~tK0.\
^ \ Co Com vti'ia. C"il O
^Z-2-S **
«ul ov\\v^^
~Ci_o o-U> 1 CL M. «_C.\-1
c_ i'AC.c.l
-------
PRELIMINARY CAPITAL
D-4
t=ttt—
OP
^>^>.(5\u.v ^W. V3g.\o>_ij wc\ >gP* v-€»
»WA«
C r nuusCR
D>c> ^gi\g.<.,
o»re
vlc-H • Zl. )o,l9
JOSMUttfttA
9\0 I - £
itcTiOx Muua(a
ItCIOB MUMld 4 Mla(
name of facility! o s s c r i p t i o nj quant.) •
\.
PuP^U/,1 <4\ ;>~ 5 1 to S-twkMl«<^
A' <£>
1
/
>
-------
PRELIMINARY CAPITAL
D-5
s-Jlf
OF
PffOJCCT
1"* I .-\A <•- - J. .A /. 91
Gr*-ste+ ggr tit,"(ZAlfJ. 7T--<.sJ
Bet il+- A *Jd 'At* " C » **r
'G£J% / "
Vr V " /a/ygt
*
-~~~<£TT6. f
J
-------
PRELIMINARY CAPITAL
FTTT
D-6
OF
*»OJ«CT
Cj Lco1*^ Sw^Tt?»V\.
PHAS5
6 p NUMaen
ST
oa re
;08
^z*, iqi^t
9\ot- F
CAse Nu*aeR
J€CTlQ* NU*&Cft
NAME OF FACILITY,'0 5 SCR I PT1 0 N-CUANV.
¦ 11
"Pi ~P
trt<7
Saw pk Pug. 2o' * '"<6
Ca.f \aow ^tg.a.1
~
/vr. *<*
3 • / "& Ja
-------
PRELIMINARY CAPITAL
D-7
i-in
OF
PPOJtCT
ClQS£t^> VE vi T *^>N *?Tn TV\ TWg. ^.r*v\S * OVN *
phase
E P NU«8£»
"O'Gp- I£tcik<,oisJ
OaTC
J*m. 2.3. '9iq
JOS HU-it8
^I0\-F
case HUMftiR
SECTION NUMBER
name OF FACILITY('DESCRIPTION! OUANi". 11
1
I.
t K
P» V>\ VN v 4oo' L£,
¦)
-------
APPENDIX E
LIST OF EPA INFORMATION SOURCES
-------
E-3
LIST OF EPA INFORMATION SOURCES*
1. J. W. Blackburn, Hydroscience, Inc., Trip Reports for Visit to Allied Chemical Co.,
Hopewell, VA, on the Adipic Acid Process, the Cyclohexanol/Cyclohexanone Process,
and the Caprolactum Process, Feb. 21, 1978.
2. D. H. Bolb, CY-RO Industries, letter to R. E. White, Hydroscience, Inc., with
information on American Cyanamid's Kenner, LA, methyl methacrylate process,
May 4, 1978.
3. J. L. Lawson, Hydroscience, Inc., Trip Report for Visit to Amoco Chemicals,
Joliet, IL, on the Maleic Anhydride Process, Jan. 24, 1978.
4. S. W. Dylewski, Hydroscience, Inc., Trip Reports for Visit to AMOCO, Decatur, AL,
on the Dimethyl Terephthalate Process, Oct. 31, 1977.
5. R. L. Standifer, Hydroscience, Inc., Trip Reports for Visit to ARCO, Channelview, TX,
on the Butadiene Process and the Ethylene Process, Aug. 16, 1977.
6. J. L. Lawson, Hydroscience, Inc., Trip Report for Visit to BASF, Geismar, LA,
on the Ethylene Oxide Process, July 12, 1977.
7. J. L. Lawson, Hydroscience, Inc., Trip Report for Visit to Borden, Fayettville, NC,
on the Formaldehyde Process, Aug. 24, 1977.
8. S. W. Dylewski, Hydroscience, Inc., Trip Reports for Visit to Carolina Eastman,
Columbia, SC, on the Dimethyl Terephthalate Process and on the Terephthalic Acid
Process, Dec. 6, 1977.
9. J. C. Cudahy, Hydroscience, Inc., Trip Report for Visit to Celanese, Clear Lake, TX,
on the Acetaldehyde Process, Sept. 22, 1977.
10. J. A. Key, Hydroscience, Inc., Trip Report for Visit to Celanese, Clear Lake, TX,
on the Acetic Acid Process, Oct. 12, 1977.
11. R. W. Helsel, Hydroscience, Inc., Trip Reports for Visit to Celanese, Pampa, TX,
on the Acetic Anhydride Process and on the Ethyl Acetate Process, Mar. 1, 1978.
12. J. L. Lawson, Hydroscience, Inc., Trip Report for Visit to Celanese, Clear Lake, TX,
on the Ethylene Oxide Process, June 21, 1977.
13. J. L. Lawson, Hydroscience, Inc., Trip Report for Visit to Celanese, Bishop, TX,
on the Formaldehyde Process, July 26, 1977.
14. J. A. Key, Hydroscience, Inc., Trip Report for Visit to Celanese, Bishop, TX,
on the Methanol Process, Oct.11, 1977.
15. S. W. Dylewski, Hydroscience, Inc., Trip Report for Visit to Celanese, Bay City, TX,
on the Vinyl Acetate Process, Sept. 28, 1977.
16. J. A. Key, Hydroscience, Inc., Trip Report for Visit to Cosmar, Carville, LA,
on the Ethylbenzene/Styrene Processes, June 16, 1977.
*All sources on file at EPA, ESED, Research Triangle Park, NC.
-------
17.
18.
19.
20.
21.
22.
23.
24,
25
26
27
28
29
30
31
32
33
34
E-4
J. L. Lawson, Hydroscience, Inc., Trip Report for Visit to Denka (Petrotex),
Houston, TX, on the Maleic Anhydride Process, Nov. 11, 1977.
J. A. Key, Hydroscience, Inc., Trip Report for Visit to Dow Chemical, Freeport, TX,
on the Ethylbenzene/Styrene Process, July 28, 1977.
J. W. Blackburn, Hydroscience, Inc., Trip Reports for Visit to Dupont, Memphis, TN,
on the Acetone Cyanohydrin Process and on the Methyl Methacrylate Process,
Jan. 10, 1978.
J. A. Key, Hydroscience, Inc., Trip Report for Visit to Dupont, Beaumont, TX,
on the Acrylonitrile Process, Sept. 7, 1977.
C. W. Stuewe, Hydroscience, Inc., Trip Report for Visit to Dupont, Beaumont, TX,
on the Aniline/Nitrobenzene Process, Sept. 7, 1977.
J. W. Blackburn, Hydroscience, Inc., Trip Report for Visit to Exxon, Baytovm, TX,
on the Cvclohexane Process, Sept. 15, 1977.
R. L. Standifer, Hydroscience, Inc., Trip Report for Visit to Gulf Oil Co.,
Cedar Bayou, TX, on the Ethylene Process, Sept. 13, 1977.
S. W. Dylewski, Hydroscience, Inc., Trip Report for Visit to Hereofina, Wilmington,
NC, on the Dimethyl Terephthalate Process, Nov. 17, 1977.
C. A. Peterson, Hydroscience, Inc., Trip Report for Visit to Monsanto, Sauget, IL,
on the Chlorobenzene Process, Jan. 27, 1978.
J. L. Lawson, Hydroscience, Inc., Trip Report for Visit to Monsanto, St. Louis, MO,
on the Maleic Anhydride Process, Oct. 20, 1977.
J. A. Key, Hydroscience, Inc., Trip Reports for Visit to Monsanto, Texas City, TX,
on the Acetic Acid Process and on the Methanol Process, Dec. 13, 1977.
C. A. Peterson, Hydroscience, Inc., Trip Report for Visit to Monsanto, Alvin, TX,
on the Alkyl Benzene Process, Nov. 8, 1977.
C. W. Stuewe, Hydroscience, Inc., Trip Report for Visit to Monsanto, Alvin, TX,
on the Phenol/Acetone Processes, July 28, 1977.
J. W. Blackburn, Hydroscience, Inc., Trip Report for Visit to Monsanto, Pensacola, FL*
on the Cyclohexanol/Cyclohexanone Process, Feb. 8, 1978.
W. D. Bruce, Hydroscience, Inc., Trip Report for Visit to Nipro, Augusta, GA,
on the Caprolactam Process, Apr. 18, 1978.
J. W. Blackburn, Hydroscience, Inc., Trip Report for Visit to Phillips, Puerto Rico,
on the Cyclohexane Process, Sept. 20, 1977.
S. W. Dylewski, Hydroscience, Inc., Trip Report for Visit to PPG, New Martinsville,
WV, on the Chlorobenzene Process, Sept. 7, 1977.
J. L. Lawson, Hydroscience, Inc., Trip Report for Visit to Reichhold, Morris, IL,
on the Maleic Anhydride Process, July 28, 1977.
-------
E-5
35. J. W. Blackburn, Hydroscience, Inc., Trip Reports for Visit to Rohm and Haas,
Deer Park, TX, on the Acrylic Acid Process and on the Methyl Methacrylate Process,
Nov. 1, 1977.
36. C. W. Stuewe, Hydroscience, Inc., Trip Report for Visit to Rubicon, Geismar, LA,
on the Aniline/Nitrobenzene Processes, July 19, 1977.
^37. C. A. Peterson, Hydroscience, Inc., Trip Report for Visit to Shell, Norco, LA,
on the Acrolein/Glycerine Processes, Jan. 25, 1978.
»38. R. E. Van Ingen, Shell Oil Co., Houston, TX, letter to Don Goodwin, EPA, ESED,
with information on Shell's vinyl chloride monomer process, Dec. 6, 1974.
39. J. W. Blackburn, Hydroscience, Trip Reports for Visit to Union Carbide, Taft, LA,
on the Acrylic Acid Process, the Ethyl Acrylate Process, and the Heavy Acrylic
Esters Process, Dec. 8, 1977.
40. C. A. Peterson, Hydroscience, Inc., Trip Report for Visit to Union Carbide,
Institute, WV, on the Linear Alkylbenzene Process, Dec. 7, 1977.
41. J. L. Lawson, Hydroscience, Inc., Trip Report for Visit to Union Carbide,
S. Charleston, WV, on the Ethylene Oxide, Glycol Processes, Dec. 7, 1977.
42. J. L. Lawson, Hydroscience, Inc., Trip Report for Visit to Union Carbide,
Sea Drift, TX, on the Glycol Ethers Process, Feb. 14, 1978.
43. J. A. Key, Hydroscience, Inc., Trip Report for Visit to Vistron, Lima, OH,
on the Acrylonitrile Process, Oct. 4, 1977.
44. Phenol process as reported by Pullman Kellogg, Apr. 28, 1978.
45. Ethylene process as.reported by Pullman Kellogg, Apr. 4, 1978.
46. Butadiene process as reported by Pullman Kellogg, Apr. 4, 1978.
47. Urea Process as reported by Pullman Kellogg, May 11, 1978.
48. Methanol Process as reported by Pullman Kellogg, May 23, 1978.
£9. Cumene Process as reported by Pullman Kellogg, May 26, 1978.
50. Benzene Process as reported by Pullman Kellogg, May 11, 1978.
ft
51. Cyclohexane Process as reported by Pullman Kellogg, Apr. 17, 1978.
52. Formaldehyde Process, as reported by Walk, Haydel and Associates, June 27, 1978.
------- |